Service is a concept that separates the concerns of an organization into (1) the value created for users and (2) the way the organization manages its resources to provide this value. The discipline of management of information technology (IT) uses services to coordinate and to optimize the use of IT resources (servers, applications, databases, etc.) in a way that brings value to users.

The concrete application of the service concept is challenging due to its abstract, interdependent and recursive nature. We experienced this challenge while collaborating with the IT department of our university (École Polytechnique Fédérale de Lausanne, EPFL) when the IT department adopted the IT Infrastructure Library (ITIL) best-practices framework for IT service management. As researchers, we have the goal of improving the understanding of services as a means to structuring what people and organizations do. In the context of the IT department, we studied how to apply the service concept internally within the IT department, and externally (as business services) in the overall organization.

In this thesis, we model services by using systems thinking principles. In particular, we use and improve SEAM, the systemic service-modeling method developed in our laboratory. Our main result is an ontology for SEAM service modeling. Our contributions are the heuristics that define how the ontology relates to a perceived reality: for example, the heuristics focus on behavior rather than organization and they put an emphasis on service instances rather than service types. We also define alignment between service systems, based on the properties of the systems¿ behavior.

We show how to model an organization by implementing the concept of service as defined by our ontology. This ontology supports the design of service systems that align across both IT and business services. During our work with over one hundred IT services, we developed several visualization prototypes of a service cartography; we use these prototypes to describe and to relate the different views required for managing services.

Our results offer a concrete way to implement the abstract concept of services. This way could be of interest for any organization willing to embark on a large-scale service project.